graphite-dev team mailing list archive

[chrismd <question83117@xxxxxxxxxxxxxxxxxxxxx>]

Question #83117 on Graphite changed:
https://answers.launchpad.net/graphite/+question/83117

    Status: Open => Answered

chrismd proposed the following answer:
It's really a matter of preference, but I find that volatile parts of a
metric name (by which I mean, parts that are likely to change or have
dynamic values) should be kept deep in the hierarchy to make navigation
cleaner.

You can organize your data however you would like but I would highly
recommend following this method: Each level in a particular hierarchy
should have a well-defined and consistent meaning. The reason is so that
using '*' lets you make selections of related metrics more easily.

A good approach is to consider the effect of using '*' at each level of
your hierarchy and ask yourself if that yields a useful set of paths or
would yield a confusing or unrelated set of paths. Note that this
doesn't mean that your entire hierarchy has to follow a single scheme,
but rather at place in the hierarchy there should be a meaningful scheme
above it that it fits into and a meaningful scheme below it that makes
sense for whatever the current path is referring to.

For example I like to use the following scheme lately:

applications.<app name>.<server name>.<scheme that makes sense for the given app>
server.<server name>.<category>.<metric>
urls.<site name>.<page name>.<metric>

The idea being that for any given metric, its path reflects what it is
with a minimal amount of "fields" that are ordered from most generic to
most specific. I hope this helps.

-- 
You received this question notification because you are a member of
graphite-dev, which is an answer contact for Graphite.